Hypothalamic supramammillary neurons that project to the medial 1 septum control wakefulness

Abstract


Introduction 35
The supramammillary nucleus (SuM) is a hypothalamic region lying above the mammillary 36 body, and provides abundant projections to numerous brain regions like the hippocampus, 37 septum, frontal cortex, and cingulate cortex (Pan & McNaughton, 2004;Vertes, 1992). 38 Recent advances in high-performance recording and manipulation techniques have enabled 39 extensive studies of SuM functions, subsequently revealing its involvement in numerous 40 processes such as episodic memory Qin et al., 2022), novelty detection (Chen 41 et al., 2020), theta rhythm (Billwiller et al., 2020;Kocsis & Vertes, 1994), locomotion (Farrell 42 et al., 2021), hippocampal neurogenesis (Li et al., 2022), and wakefulness (Pedersen et al.,43 2017). In particular, one previous study demonstrated that SuM glutamatergic neurons serve 44 as a key node for arousal, and chemogenetic activation of SuM glutamatergic neurons, but 45 not GABAergic neurons, produces sustained arousal (Pedersen et al., 2017). However, which 46 downstream brain regions are involved in the SuM control of arousal remains unknown. 47 The medial septum (MS), which primarily contains cholinergic, GABAergic and 48 glutamatergic neurons (Hajszan et al., 2004;Kiss et al., 1990), has been suggested to mediate 49 different brain functions like locomotion (Fuhrmann et al., 2015), learning and memory 50 (Boyce et al., 2016;Lecourtier et al., 2011), hippocampal theta generation (Buzsáki, 2002), 51 and wakefulness Osborne, 1994). Among these functions, MS glutamatergic 52 neurons were shown to control wakefulness by activating lateral hypothalamic glutamatergic 53 neurons . Furthermore, a recent study has demonstrated that SuM 54 glutamatergic neurons project to MS glutamatergic neurons and are responsible for 55 modulating the motivation for environmental interaction (Kesner et al., 2021). Based on this 56 established anatomical connection and combined findings, we hypothesized that a SuM-MS 57 projection may control wakefulness. 58 To test this hypothesis, we performed circuit-specific optical Ca 2+ and optrode 59 recordings in SuM-MS projection across sleep-wakefulness cycles. We identified a set of 60 wake-active neurons in SuM that project to MS. Optogenetic or chemogenetic activation of 61 5 12.4 ± 2.1 Hz; REM, 10.6 ± 2.6 Hz; NREM, 6.6 ± 1.7 Hz; Friedman's ANOVA and Wilcoxon 122 signed-rank tests, n = 23 neurons from 8 mice, wakefulness versus NREM, P = 0.001, REM 123 versus NREM, P = 0.0002, wakefulness versus REM, P = 0.24). 124 Analysis of firing modulation by SuM MS projecting neurons during these three states 125 followed by calculation of firing rates in wake-active neurons during state transitions ( Figure  126 2I) revealed that these wake-active neurons had higher firing rates during 127 NREM-wakefulness or REM-wakefulness transitions, but lower firing rates during 128 wakefulness-NREM transitions ( Figure 2J). These results established that wake-active 129 neurons were indeed present in SuM-MS projection, likely contributing to control of 130 wakefulness. 131

Optrode construction for in vivo recording
The optical fiber was fixed to tetrodes with the tips being ~500 μm shorter than the tetrode 318 tips. The light from a laser diode (450 nm) was collimated to the optical fiber at the opposite 319 15 end with a maximal light intensity measured by an optical power meter (PM100D, Thorlabs). 320 Optical adhesive was used to connect the laser diode and optical fiber. 321

Surgical procedures 322
For all surgeries, mice were anesthetized with 3% isoflurane in oxygen for 3-5 min and then 323 placed into a stereotaxic frame with an isoflurane concentration maintained at 1%-2%. A 324 heating pad was put under the mice to maintain a temperature of ~37 °C throughout the 325 surgery process. After surgery, the mice were placed back in warm cages and allowed to fully 326 recover. Moreover, they received one dose of dexamethasone sodium phosphate (1mg/ml, 327 0.1ml/10g/d) and ceftriaxone sodium (50mg/ml, 0.1ml/10g/d) per day by intraperitoneal 328 injection for 3 consecutive days to reduce inflammation (Li et  were used. The prepared fiber probe was inserted through a small cranial window above MS 348 (AP: 1.0 mm, ML: 0.5 mm, 5° angle towards the midline) to a depth of 3.5 mm. Blue 349 16 light-curing dental cement (595989WW, Tetric EvoFlow) was applied to fix the probe to the 350 skull. Further reinforcement was achieved with a common dental cement mixture in super 351 glue. 352 For optrode implantation, mice expressing ChR2 in SuM MS projecting neurons were used. 353 Similarly, the previously described optrode was inserted after a craniotomy above SuM was 354 made. The implantation depth was 4.7 mm from the dura. After a full recovery (the body 355 weight started to increase), the optrode was gradually advanced to the target depth of ~5.0 356 mm by micro-drive. 357 For EEG-EMG electrodes implantation, three EEG electrodes made by stainless steel 358 screws were inserted into the craniotomy holes, with two above the frontal lobe (AP: 1.3 359 mm, ML: ± 1.2 mm) and the third one above the parietal lobe (AP: -3.2 mm, ML: 3.0 mm). 360 Two fine-wire EMG electrodes were inserted into the neck musculature for EMG recording. 361 Before all recording and manipulation experiments, mice were connected to optical and 362 electrophysiological recording cables in the recording cages to habituate for 3 consecutive 363 days. 364

Fiber recording 365
A previously described fiber photometry system was used for Ca 2+ recording (Qin et al., 2018;366 Qin et al., 2022). The recording was performed in jGCaMP7b-expressing mice with a fiber 367 probe implanted in MS. Ca 2+ activity (2 KHz), EEG-EMG signals (200 Hz), and behavioral 368 videos (25 Hz) were simultaneously recorded across sleep-wakefulness cycles. Offline event 369 makers were used to synchronize these three forms of signals. 370

Optrode recording 371
Excitation light pulses (450 nm wavelength, 10 ms duration, ~10 mW intensity, 0.5 s interval) 372 were applied in optrode-implanting mice to identify SuM MS projecting neurons. Units evoked 373 by light stimulation with short spike latency (< 8 ms for all the units in our data) and high 374 response reliabilities (> 73% for all the units in our data) were identified as ChR2-positive 375 neurons. Then electrophysiological recordings (sampled at 20 KHz), EEG-EMG recording, and 376 behavioral video recordings were simultaneously conducted across sleep-wakefulness cycles 377 in the light phase. After all recordings were finished, an electrical lesion (current with 30 µA 378 intensity and 12 s duration) was made to verify the recording sites. 379